Mind-to-mind heteroclinic coordination: Model of sequential episodic memory initiation.

Retrieval of episodic memory is a dynamical process in the large scale brain networks. In social groups, the neural patterns, associated with specific events directly experienced by single members, are encoded, recalled, and shared by all participants. Here, we construct and study the dynamical model for the formation and maintaining of episodic memory in small ensembles of interacting minds. We prove that the unconventional dynamical attractor of this process-the nonsmooth heteroclinic torus-is structurally stable within the Lotka-Volterra-like sets of equations. Dynamics on this torus combines the absence of chaos with asymptotic instability of every separate trajectory; its adequate quantitative characteristics are length-related Lyapunov exponents. Variation of the coupling strength between the participants results in different types of sequential switching between metastable states; we interpret them as stages in formation and modification of the episodic memory.

Relating the sequential dynamics of excitatory neural networks to synaptic cellular automata.

We have developed a new approach for the description of sequential dynamics of excitatory neural networks. Our approach is based on the dynamics of synapses possessing the short-term plasticity property. We suggest a model of such synapses in the form of a second-order system of nonlinear ODEs. In the framework of the model two types of responses are realized-the fast and the slow ones. Under some relations between their timescales a cellular automaton (CA) on the graph of connections is constructed. Such a CA has only a finite number of attractors and all of them are periodic orbits. The attractors of the CA determine the regimes of sequential dynamics of the original neural network, i.e., itineraries along the network and the times of successive firing of neurons in the form of bunches of spikes. We illustrate our approach on the example of a Morris-Lecar neural network.

On the origin of reproducible sequential activity in neural circuits.

Robustness and reproducibility of sequential spatio-temporal responses is an essential feature of many neural circuits in sensory and motor systems of animals. The most common mathematical images of dynamical regimes in neural systems are fixed points, limit cycles, chaotic attractors, and continuous attractors (attractive manifolds of neutrally stable fixed points). These are not suitable for the description of reproducible transient sequential neural dynamics. In this paper we present the concept of a stable heteroclinic sequence (SHS), which is not an attractor. SHS opens the way for understanding and modeling of transient sequential activity in neural circuits. We show that this new mathematical object can be used to describe robust and reproducible sequential neural dynamics. Using the framework of a generalized high-dimensional Lotka-Volterra model, that describes the dynamics of firing rates in an inhibitory network, we present analytical results on the existence of the SHS in the phase space of the network. With the help of numerical simulations we confirm its robustness in presence of noise in spite of the transient nature of the corresponding trajectories. Finally, by referring to several recent neurobiological experiments, we discuss possible applications of this new concept to several problems in neuroscience.

Space-time complexity in Hamiltonian dynamics.

New notions of the complexity function C(epsilon;t,s) and entropy function S(epsilon;t,s) are introduced to describe systems with nonzero or zero Lyapunov exponents or systems that exhibit strong intermittent behavior with "flights," trappings, weak mixing, etc. The important part of the new notions is the first appearance of epsilon-separation of initially close trajectories. The complexity function is similar to the propagator p(t(0),x(0);t,x) with a replacement of x by the natural lengths s of trajectories, and its introduction does not assume of the space-time independence in the process of evolution of the system. A special stress is done on the choice of variables and the replacement t-->eta=ln t, s-->xi=ln s makes it possible to consider time-algebraic and space-algebraic complexity and some mixed cases. It is shown that for typical cases the entropy function S(epsilon;xi,eta) possesses invariants (alpha,beta) that describe the fractal dimensions of the space-time structures of trajectories. The invariants (alpha,beta) can be linked to the transport properties of the system, from one side, and to the Riemann invariants for simple waves, from the other side. This analog provides a new meaning for the transport exponent mu that can be considered as the speed of a Riemann wave in the log-phase space of the log-space-time variables. Some other applications of new notions are considered and numerical examples are presented.

Generalized synchronization of chaos in noninvertible maps.

The properties of functional relation between a noninvertible chaotic drive and a response map in the regime of generalized synchronization of chaos are studied. It is shown that despite a very fuzzy image of the relation between the current states of the maps, the functional relation becomes apparent when a sufficient interval of driving trajectory is taken into account. This paper develops a theoretical framework of such functional relation and illustrates the main theoretical conclusions using numerical simulations.

A class of cellular automata modeling winnerless competition.

Neural units introduced by Rabinovich et al. ("Sensory coding with dynamically competitive networks," UCSD and CIT, February 1999) motivate a class of cellular automata (CA) where spatio-temporal encoding is feasible. The spatio-temporal information capacity of a CA is estimated by the information capacity of the attractor set, which happens to be finitely specified. Two-dimensional CA are studied in detail. An example is given for which the attractor is not a subshift. (c) 2002 American Institute of Physics.

Multivalued mappings in generalized chaos synchronization.

The onset of generalized synchronization of chaos in directionally coupled systems corresponds to the formation of a continuous mapping that enables one to persistently define the state of the response system from the trajectory of the drive system. A recently developed theory of generalized synchronization of chaos deals only with the case where this synchronization mapping is a single-valued function. In this paper, we explore generalized synchronization in a regime where the synchronization mapping can become a multivalued function. Specifically, we study the properties of the multivalued mapping that occurs between the drive and response systems when the systems are synchronized with a frequency ratio other than one-to-one, and address the issues of the existence and continuity of such mappings. The basic theoretical framework underlying the considered synchronization regimes is then developed.

Pesin's dimension for Poincare recurrences.

A new characteristic of Poincare recurrences is introduced. It describes an average return time in the framework of a general construction for dimension-like characteristics. Some examples are considered including rotations on the circle and the Denjoy example. (c) 1997 American Institute of Physics.

Statistical properties of 2-D generalized hyperbolic attractors.

Recently Pesin introduced a large class of hyperbolic attractors, and for those attractors he established the Smale spectral decomposition. In this paper our main results are a stretched exponential bound on the decay of correlations and the central limit theorem. Also we will obtain conditions under which two well known attractors-those of Belykh and Lozi-are subject to our main results. (c) 1995 American Institute of Physics.

Traveling waves in lattice models of multidimensional and multicomponent media. II. Ergodic properties and dimension.

We demonstrate a spatio-temporal chaos in lattice models of multidimensional and multicomponent media on the set of traveling waves solutions running with large enough velocities. We describe stability properties of such solutions, construct invariant measures with "good" ergodic properties concentrated on the above set and study different types of dimensions including the correlation dimension.